Sol_Badguy
New Member
- Joined
- Feb 2, 2025
- Messages
- 21 (0.42/day)
Hello,
First of all I want to apologize if this subject has already been discussed, if so please point me in the direction of that thread and ignore what's below.
Okay so the issue I'm talking about is regarding the percentages we see in these charts, and how do these percentages reflect (or not) what's stated in the chart header.
www.techpowerup.com
www.techpowerup.com
The following is stated: Positive number: CPU_A Faster, Negative number: CPU_B Faster
This wording denotes symmetry.
In the tests we have two categories of results/scores, depending on the application:
1. Higher results/scores are better
2. Lower results/scores are better
The formulas used for the positive numbers are the following:
1. For results where higher is better:
(score_A - score_B) / score_B * 100 = % indicating how much stronger CPU_A is in relation to CPU_B
2. For results where lower is better:
(score_B - score_A) / score_B * 100 = % indicating how much faster CPU_A is in relation to CPU_B
These percentages match the first statement: Positive number: CPU_A Faster
The formulas used for the negative numbers are the following:
1. For results where higher is better:
(score_A - score_B) / score_B * 100 = % indicating how much weaker CPU_A is in relation to CPU_B
2. For results where lower is better:
(score_B - score_A) / score_B * 100 = % indicating how much slower CPU_A is in relation to CPU_B
These percentages DON'T match the second statement: Negative number: CPU_B Faster
They actually denote how much weaker/slower CPU_A is in relation to CPU_B which in a roundabout way gets to the same conclusion but the meaning of the words and also how the result is presented is different because the approach is asymmetrical.
Concrete example:
The percentages are reversed, because CPU_A is always used as a baseline, as a result the negative percentages for higher is better are decreased and those for lower is better are increased.
Sure we could say that in the end they even out, but I don't consider this approach correct because it's pottentially confusing and also misleading by changing the order of the percentage bars.
I understand that the CPU that's being reviewed gets the spotlight and is used as a baseline (CPU_A in this case), but when CPU_B gets a better result I want the percentage to reflect that, how much better CPU_B is in relation to CPU_A and not how much slower CPU_A is in relation to CPU_B.
As I said it could be confusing for some people, they read the chart header, then look at the negative percentages and possibly misinterpret them.
www.techpowerup.com
First of all I want to apologize if this subject has already been discussed, if so please point me in the direction of that thread and ignore what's below.
Okay so the issue I'm talking about is regarding the percentages we see in these charts, and how do these percentages reflect (or not) what's stated in the chart header.
AMD Ryzen 9 9950X3D Review - Great for Gaming and Productivity
The new AMD Ryzen 9 9950X3D brings Zen 5 with 3D V-Cache to the high-end. This new $700 flagship offers the best application performance, beating even the 9950X, and at the same time you get a fantastic gaming experience that's better than any other non-X3D processor on the market.
AMD Ryzen 7 9800X3D Review - The Best Gaming Processor
The Ryzen 7 9800X3D establishes AMD as the leader in gaming performance. This Zen 5-based X3D chip is not only fast, it also comes with full support for overclocking. Besides gaming, application performance is considerably improved over the 7800X3D, but that comes at a price.
This wording denotes symmetry.
In the tests we have two categories of results/scores, depending on the application:
1. Higher results/scores are better
2. Lower results/scores are better
The formulas used for the positive numbers are the following:
1. For results where higher is better:
(score_A - score_B) / score_B * 100 = % indicating how much stronger CPU_A is in relation to CPU_B
2. For results where lower is better:
(score_B - score_A) / score_B * 100 = % indicating how much faster CPU_A is in relation to CPU_B
These percentages match the first statement: Positive number: CPU_A Faster
The formulas used for the negative numbers are the following:
1. For results where higher is better:
(score_A - score_B) / score_B * 100 = % indicating how much weaker CPU_A is in relation to CPU_B
2. For results where lower is better:
(score_B - score_A) / score_B * 100 = % indicating how much slower CPU_A is in relation to CPU_B
These percentages DON'T match the second statement: Negative number: CPU_B Faster
They actually denote how much weaker/slower CPU_A is in relation to CPU_B which in a roundabout way gets to the same conclusion but the meaning of the words and also how the result is presented is different because the approach is asymmetrical.
Concrete example:
Test 1: higher is better
score_A: 150
score_B: 100
Using the formula: (score_A - score_B) / score_B * 100 = %
We get that CPU_A is 50 % stronger than CPU_B.
Test 2: lower is better
score_A: 4
score_B: 6
Using the formula: (score_B - score_A) / score_B * 100 = %
We get that CPU_A is 33 % faster than CPU_B.
Test 3: higher is better
score_A: 100
score_B: 150
Using the formula: (score_A - score_B) / score_B * 100 = %
We get that CPU_A is -33 % weaker than CPU_B.
Test 4: lower is better
score_A: 6
score_B: 4
Using the formula: (score_B - score_A) / score_B * 100 = %
We get that CPU_A is -50 % slower than CPU_B.
score_A: 150
score_B: 100
Using the formula: (score_A - score_B) / score_B * 100 = %
We get that CPU_A is 50 % stronger than CPU_B.
Test 2: lower is better
score_A: 4
score_B: 6
Using the formula: (score_B - score_A) / score_B * 100 = %
We get that CPU_A is 33 % faster than CPU_B.
Test 3: higher is better
score_A: 100
score_B: 150
Using the formula: (score_A - score_B) / score_B * 100 = %
We get that CPU_A is -33 % weaker than CPU_B.
Test 4: lower is better
score_A: 6
score_B: 4
Using the formula: (score_B - score_A) / score_B * 100 = %
We get that CPU_A is -50 % slower than CPU_B.
The percentages are reversed, because CPU_A is always used as a baseline, as a result the negative percentages for higher is better are decreased and those for lower is better are increased.
Sure we could say that in the end they even out, but I don't consider this approach correct because it's pottentially confusing and also misleading by changing the order of the percentage bars.
I understand that the CPU that's being reviewed gets the spotlight and is used as a baseline (CPU_A in this case), but when CPU_B gets a better result I want the percentage to reflect that, how much better CPU_B is in relation to CPU_A and not how much slower CPU_A is in relation to CPU_B.
As I said it could be confusing for some people, they read the chart header, then look at the negative percentages and possibly misinterpret them.
Performance ranking oddity
I was comparing 2080 Ti (my current card) to 9070 and 9070 XT. Perhaps im simply reading percentages wrong but: When selecting 2080 Ti as baseline 100%, then 9070 is 69% faster and 9070 XT is 87% faster. The cap between the 9070 models is 18%. However when selecting 9070 as baseline 100% then...
www.techpowerup.com